An Episomal Expression Vector for Screening Mutant Gene Libraries in Pichia Pastoris

Home , Expression vector

Plasmid 54 (2005) 80–85 www.elsevier.com/locate/yplas Short communication An episomal expression vector for screening mutant gene libraries in Pichia pastoris

Charles C. Lee*, Tina G. Williams, Dominic W.S. Wong, George H. Robertson

USDA-ARS-WRRC, 800 Buchanan St., Albany, CA 94710, USA

Received 7 October 2004, revised 6 December 2004 Available online 20 January 2005 Communicated by Manuel Espinosa

Abstract

Screening mutant gene libraries for isolating improved enzyme variants is a powerful technique that benefits from effective and reliable biological expression systems. Pichia pastoris is a very useful organism to express proteins that are inactive in other hosts such as Escherichia coli and Saccharomyces cerevisiae. However, most P. pastoris expression plasmids are designed to integrate into the host chromosome and hence are not as amenable to high-throughput screening projects. We have designed a P. pastoris expression vector, pBGP1, incorporating an autonomous replication sequence that allows the plasmid to exist as an episomal element. This vector contains the a-factor signal sequence to direct secretion of the mutant enzymes. Expression of the genes is driven by the constitutive GAP promoter, thus eliminating the need for timed or cell density-specific inductions. The pBGP1 plasmid was used to screen a xylanase gene library to isolate higher activity mutants. Published by Elsevier Inc.

Keywords: Pichia pastoris; Episomal expression vector

1. Introduction formed with libraries of randomly mutated genes encoding the enzyme of interest (Cohen et al., There are many efforts directed at improving 2001). Clones that demonstrate an increased activ- enzymes involved in industrial processes in order ity relative to wild type can be isolated and further to increase cost and energy efficiency. One of the characterized. most promising methods to obtain better enzymes In order to screen a library for increased activ- is to screen microorganisms that have been trans- ity, it is critical to develop an effective expression system. It is desirable to have an expression vector * Corresponding author. Fax: +1 510 559 5940. that will direct secretion of the enzyme into the E-mail address: [email protected] growth media. This allows for the rapid assaying

0147-619X/$ - see front matter. Published by Elsevier Inc. doi:10.1016/j.plasmid.2004.12.001 Short communication / Plasmid 54 (2005) 80–85 81 of the enzyme activity without the need to release contrast, an episome could be isolated by a simple the protein by cell lysis or osmotic shock. Secreted plasmid preparation procedure. enzymes are often also easier to purify because of In this report, we describe the construction of a the relatively low levels of contaminating cellular new P. pastoris expression vector for screening proteins. mutant gene libraries. The plasmid contains both The most common host organisms used for bacterial and P. pastoris autonomous replication screening libraries are the bacterium Escherichia sequences that allow the vector to exist as an coli and the yeast Saccharomyces cerevisiae. E. coli episome. A zeocin-resistance gene permits selection is popular because of its rapid growth and high of the plasmid in P. pastoris. We demonstrate the transformation efficiency. However, many pro- utility of the plasmid by screening a library contain- teins, especially those of eukaryotic origin, ex- ing mutant gene variants of a xylanase enzyme that pressed in E. coli are not active due to improper was not active in either E. coli or S. cerevisiae host. folding and/or lack of necessary post-translational modifications, such as glycosylation and signal 1.1. Strains, growth media, and enzymes sequence processing. S. cerevisiae does have the cellular machinery to modify enzymes post-transl- Escherichia coli bacterial strain JM109 ationally, so this yeast often succeeds where the (Promega, Madison, WI) was cultured in Luria– bacterium fails (Dean, 1999; Strahl-Bolsinger Bertani broth (1% tryptone, 0.5% yeast extract, et al., 1999). Unfortunately, some proteins ex- and 1% sodium chloride) supplemented with pressed in S. cerevisiae are hyperglycosylated, a ampicillin (50 lg/ml) when appropriate. The condition that can also result in loss of activity P. pastoris yeast strain GS115 (Invitrogen, Carls- (Eckart and Bussineau, 1996). bad, CA) was cultured on YPD (1% yeast extract, Another host that has been used successfully 2% peptone, and 2% glucose) supplemented with when both E. coli and S. cerevisiae have failed is zeocin (100 lg/ml) when appropriate. YPDS Pichia pastoris. This methylotrophic yeast has the (YPD containing 1 M sorbitol) was used when plat- capacity to grow to very high densities and can ing yeast that were transformed by electroporation. produce large amounts of protein (Cereghino For solid media, 1.5 and 2% agar were used for the and Cregg, 2000). Although P. pastoris also post- bacterial and yeast media, respectively. translationally modifies expressed enzymes, it does All media components were manufactured by not tend to hyperglycosylate proteins as is the case Becton–Dickinson (Sparks, MD), and all enzymes with S. cerevisiae (Bretthauer and Castellino, were purchased from New England Biolabs 1999). (Beverly, MA) unless otherwise noted. One difficulty with using P. pastoris as a host to screen mutant libraries is that most P. pastoris 1.2. pBGP1 plasmid construction expression vectors are designed to integrate into the host chromosome. These vectors do not con- The first PCR was conducted using PfuTurbo tain an autonomous replication sequence and thus polymerase (Stratagene, La Jolla, CA), pBlue- cannot exist as episomal (i.e., replicative) plasmids. script-KS plasmid template (Stratagene), and the Instead, the vectors must be linearized, trans- primers BS938 and BS2958 to isolate a 2021 bp formed into P. pastoris, and then integrated into fragment containing the ampicillin-resistance gene the yeast chromosome. This feature of such vec- and the bacterial ColE1 origin of DNA replica- tors makes screening of the libraries more labori- tion. The second PCR employed PfuTurbo ous. First, linearized DNA does not transform as polymerase, pGAPZaA vector template (Invitro- efficiently as circular DNA, so there are fewer gen), and the primers GAP1 and GAP2405 to clones from the library that are available for anal- produce a 2405 bp fragment containing the ysis. In addition, when the clone is inserted into glyceraldehyde-3-phosphate dehydrogenase gene the host chromosome, it is much more difficult to (GAP) promoter, the a-factor secretion signal, extract for further analysis and manipulations. In and the zeocin-resistance gene. The products of 82 Short communication / Plasmid 54 (2005) 80–85 both PCRs were digested with SpeI and HindIII G1510A, T1606C, C2114G, add G at 897, add A restriction enzymes and ligated using T4 DNA li- at 1479, add C at 1517, add G at 2106, and remove gase to produce the pBS-GAP vector. extra G at 2099. A third PCR was conducted using PfuTurbo Primer sequences (restriction enzyme site used): polymerase, pYM8 plasmid template (Cregg et al., 1985), and the primers PARS1-FC and BS938: gcgaagcttacgtatgcactgcagctcactgcccgctttcc PARS1-RA to generate a 187 bp fragment con- agt (HindIII); taining the P. pastoris autonomous replication se- BS2958: gcgactagtgcacttttcggggaaatgtgcg (SpeI); quence (PARS1). The product of this PCR and GAP1: gcgactagtagatcttttttgtagaaatgtcttggtgtcctc pBS-GAP vector were digested with PstI and Hin- (SpeI); dIII restriction enzymes and ligated with T4 DNA GAP2405: gcgaagcttagcttgcaaattaaagccttcgagc ligase to produce the pBGP1 episomal expression (HindIII); vector. PARS1-FC: cgactgcagtcgagataagctgggggaacattcg The entire pBGP1 plasmid was sequenced and (PstI); compared to the predicted sequence. Multiple dis- PARS1-RA: ctgggatccaagcttcgacaattaatatttacttatt crepancies were detected in the region that origi- ttggtcaaccc (HindIII). nated from the pGAPZaA expression vector. To investigate this issue, the original pGAPZaA vector provided by Invitrogen was sequenced. It was 1.3. Features of the pBGP1 expression vector discovered that the discrepancies were present in the pGAPZaA vector. Thus, there are errors in The pBGP1 plasmid was designed with several the pGAPZaA vector sequence data file available features to serve as an episomal expression plas- on the Invitrogen web site (www.invitrogen.com). mid (Fig. 1). The P. pastoris autonomous replicat- These errors do not occur in any protein-coding ing sequence (PARS1) permitted the maintenance sequences and do not impact the functionality of of the plasmid without chromosomal integration either the pGAPZaA or the pBGP1 expression in the yeast (Cregg et al., 1985). The bacterial vectors. Corrections that should be incorporated colE1 origin of DNA replication was also included into the pGAPZaA vector sequence file are as fol- to allow the maintenance of the plasmid in E. coli. lows: T1348C, A1493G, G1494A, A1509G,

Fig. 1. pBGP1 expression vector. pGAP, GAP promoter; a-factor, secretion signal peptide; MCS, multiple cloning site derived from pGAPZaA plasmid; zeoR, zeocin-resistance gene; PARS1, P. pastoris autonomous replication sequence; colE1, bacterial replication sequence; and ampR, ampicillin-resistance gene. Short communication / Plasmid 54 (2005) 80–85 83

A signal sequence derived from the S. cerevisiae E. coli, we chose to include the ampicillin-resis- a-factor gene was included upstream of the multi- tance gene (ampR) because the ampicillin antibi- ple cloning site (MCS). Genes that are cloned in- otic is more stable and economical than zeocin. frame to this sequence will code for enzymes that The inclusion of the ampicillin-resistance gene is are secreted extracellularly into the media. Cahill another feature of pBGP1 that is distinct from and colleagues also developed a series of P. pasto- the vectors designed by Cahill and colleagues ris episomal expression vectors (Lueking et al., (Lueking et al., 2000, 2003). 2000, 2003). However, their expression vectors did not contain a signal sequence. This necessi- 1.4. Use of episomal pBGP1 to express xylanase tated a lysis procedure to release the enzymes in enzyme order to assay the activities. Expression of the genes cloned into pBGP1 is We tested the ability of the pBGP1 plasmid to driven by the GAP promoter that expresses the express a xylanase gene (xyn11A) that we had pre- gene of interest constitutively. This feature is in viously isolated from Lentinula edodes (GenBank contrast to the inducible promoters employed by Accession No. AF411252)(Lee et al., 2005). Previ- Cahill and colleagues to design their episomal ous attempts to produce the enzyme in E. coli and expression vectors. They used the AOX (methanol S. cerevisiae failed to produce protein that was ac- inducible) (Lueking et al., 2000) and the CUP (cop- tive (data not shown). When the gene was cloned per inducible) (Lueking et al., 2003) promoters. into pGAPZaA (Invitrogen), an integrating P. pas- The GAP, AOX, and CUP promoters have com- toris expression vector, active enzyme was ex- parable activities (Koller et al., 2000; Waterham pressed (Lee et al., 2005). We cloned the gene et al., 1997). However, the use of the AOX and into pBGP1 to express and secrete active Xyn11A CUP inducible promoters introduces additional from an episomal vector. Fig. 2A illustrates that ac- steps in protein expression. For instance, cell tive Xyn11A is being secreted into the culture med- growth is monitored so that the inducing agent is ia where the dye-labeled xylan substrate is digested. added at a consistent and appropriate time point. To demonstrate that the pBGP1 constructs were These manipulations add time and cost to the over- maintained as episomal elements, we wished to iso- all procedure and make a high-throughput screen- late and analyze DNA from transformed P. pasto- ing project less convenient. Inducible promoters ris colonies on an agarose gel. The PARS1 are critical in those instances in which the enzyme sequence maintains the plasmid as a single copy is toxic to the cells. However, when the enzyme of in P. pastoris (Cregg et al., 1985). The amount of interest is tolerated by the host organism, the use DNA recovered from a yeast plasmid preparation of a constitutive promoter, such as GAP, stream- is sufficient for mutagenic PCR protocols, but is lines the screening procedures. difficult to visualize by ethidium bromide staining. Although the GAP promoter is a very strong Thus, the recovered yeast plasmid was transformed constitutive promoter, the activity of the promoter into E. coli which will maintain the plasmid at sev- can be decreased by culturing the yeast on alternate eral hundred copies (Sambrook and Russell, 2001). carbon sources, such as glycerol and methanol The plasmid which had been amplified in E. coli (Sears et al., 1998; Waterham et al., 1997). Thus, was then prepared and analyzed on an agarose if the activity of an enzyme is improved through gel (Fig. 2B). As can be seen, plasmid DNA of multiple cycles of directed evolution to such an the correct size was isolated from the clones. extent that the activity assays are saturated, it should be possible to change the carbon source 1.5. Use of pBGP1 in screening mutant gene of the culture media and continue to screen for library even higher activities. To select for the plasmid in P. pastoris, a zeo- To demonstrate the feasibility of screening for cin-resistance gene was included (zeoR). Although improved enzyme variants with the pBGP1 vector, this zeocin selection marker can also be used in we subcloned a mutant xylanase gene library into 84 Short communication / Plasmid 54 (2005) 80–85

Fig. 2. Expression of active xylanase from the episomal pBGP1 vector. (A) P. pastoris GS115 cells transformed with either vector (pBGP1) or pBGP1 expression constructs containing wild type (Wt) or mutant (Mut) xylanase genes were spotted onto YPD media containing 0.1% azo-arabinoxylan. (B) Plasmid DNA was isolated (RPM Yeast Plasmid Isolation Kit, Qbiogene, Carlsbad, CA) from individual P. pastoris GS115 clones that were transformed with various pBGP1 constructs. The DNA was transformed into JM109 bacteria, and plasmid was isolated from minipreparations. The DNA was then digested with EcoRI and NotI, and the products were analyzed on an agarose gel. Predicted sizes of pBGP1 vector and xylanase gene are 4607 and 849 base pairs, respectively. M, 1 kb DNA ladder (New England Biolabs); V, pBGP1 vector; Wt, wild type pBGP1-Xyn construct; and Mut, mutant high-activity pBGP1-Xyn constructs. the vector. A pool of mutant xyn11A genes was clones of interest were inoculated into individual generated using the GeneMorph PCR mutagenesis wells of a 96-well plate containing YPD liquid kit (Stratagene), gene-specific primers, and the L. media supplemented with zeocin. After 48 h of edodes xyn11A gene. DNA shuffling was used to growth at 30 °C and 600 rpm in the Multitron further increase the diversity of the xyn11A vari- shaker (ATR, Emeryville, CA), the supernatant ants (Stemmer, 1994, 2002). Five hundred nano- was assayed for xylanase activity. Various grams of the PCR product was digested with amounts of culture supernatant were mixed with BsaJI, MnlI, or MspI. After digestion, the restric- 25 ll of 10 mg/ml azo-arabinoxylan, and enough tion enzymes were inactivated, and the individual buffer (25 mM sodium acetate, pH 5.0) was added digests were pooled. The digested DNA was then to raise the final volume to 75 ll. The reaction subjected to primerless PCR using PfuTurbo poly- was incubated at 50 °C for 10 min. Then, 150 ll merase to reassemble the full length gene (94 °C of ethanol was added to stop the reaction and for 3 min; 40 cycles of 94 °C for 1 min, 55 °C for precipitate undigested substrate. The reactions 1 min, and 72 °C for 1 min; 72 °C for 5 min). A were spun down for 30 s in a microcentrifuge, final PCR was conducted on the reassembled and 150 ll of the supernatant was placed into a DNA with gene-specific primers that had restric- chamber of a 96-well plate. The level of activity tion enzyme sites (EcoRI and NotI) located in of each clone was quantified by measuring the the linker region. This PCR product and the samples on a plate reader using a 620 nm filter. pBGP1 vector were digested with EcoRI and NotI, We screened 30,000 clones and identified several and the resulting fragments were ligated using T4 that produced larger halos (Fig. 2A) and had DNA ligase to create the mutant xyn11A gene 3-fold higher activities based on the azo-arabin- library. oxylan liquid activity assay. Plasmid DNA was The mutant xyn11A library was transformed isolated from these candidate clones and retrans- into P. pastoris GS115 yeast cells by electropora- formed into P. pastoris cells to confirm the activity tion. The transformation was spread onto 22.5 cm of the mutant genes. In addition, the plasmid square plates (Genetix, Beaverton, Oregon) con- DNA was prepared from these clones for addi- taining YPDS agar media with zeocin and 0.1% tional analyses. The plasmids were analyzed on azo-arabinoxylan from wheat (Megazyme, Wick- an agarose gel to confirm that the structure was low, Ireland). After 72 h, enzymatic activity was intact (Fig. 2B). The DNA was also sequenced to revealed by clear halos surrounding individual confirm that mutations directing amino acid colonies. Colonies that produced larger clearings, changes had occurred. These plasmids can be used or halos, were collected for further analysis. These as the starting genetic material for additional Short communication / Plasmid 54 (2005) 80–85 85 rounds of mutagenesis and screening. Thus, we Eckart, M.R., Bussineau, C.M., 1996. Quality and authenticity have demonstrated the utility of the pBGP1 epi- of heterologous proteins synthesized in yeast. Curr. Opin. somal expression vector in screening a mutant Biotechnol. 7, 525–530. Koller, A., Valesco, J., Subramani, S., 2000. The CUP1 gene library. promoter of Saccharomyces cerevisiae is inducible by copper in Pichia pastoris. Yeast 16, 651–656. Lee, C.C., Wong, D.W.S., Robertson, G.H., 2005. Cloning and Acknowledgment characterization of the Xyn11A gene from Lentinula edodes. Protein J. 24 (1), 21–26. Lueking, A., Holz, C., Gotthold, C., Lehrach, H., Cahill, D., We thank Dr. James Cregg for kindly providing 2000. A system for dual protein expression in Pichia pastoris the pYM8 plasmid. and Escherichia coli. Protein Expr. Purif. 20, 372–378. Lueking, A., Horn, S., Lehrach, H., Cahill, D.J., 2003. A dual- expression vector allowing expression in E. coli and P. References pastoris, including new modifications. Methods Mol. Biol. 205, 31–42. Bretthauer, R.K., Castellino, F.J., 1999. Glycosylation of Sambrook, J., Russell, D.W., 2001. Molecular Cloning: A Pichia pastoris-derived proteins. Biotechnol. Appl. Bio- Laboratory Manual. Cold Spring Harbor Laboratory Press, Cold Spring Harbor, NY. chem. 30, 193–200. Õ Cereghino, J.L., Cregg, J.M., 2000. Heterologous protein Sears, I.B., O Connor, J., Rossanese, O.W., Glick, B.S., 1998. A expression in the methylotrophic yeast Pichia pastoris. versatile set of vectors for constitutive and regulated gene expression in Pichia pastoris. Yeast 14, 783–790. FEMS Microbiol. Rev. 24, 45–66. Stemmer, W.P., 1994. Rapid evolution of a protein in vitro by Cohen, N., Abramov, S., Dror, Y., Freeman, A., 2001. In vitro enzyme evolution: the screening challenge of isolating the DNA shuffling. Nature 370, 389–391. one in a million. Trends Biotechnol. 19, 507–510. Stemmer, W.P., 2002. Molecular breeding of genes, pathways and genomes by DNA shuffling. J. Mol. Catal. B: Enzy- Cregg, J.M., Barringer, K.J., Hessler, A.Y., Madden, K.R., matic 19–20, 3–12. 1985. Pichia pastoris as a host system for transformations. Mol. Cell. Biol. 5, 3376–3385. Strahl-Bolsinger, S., Gentzsch, M., Tanner, W., 1999. Dean, N., 1999. Asparagine-linked glycosylation in the yeast Protein O-mannosylation. Biochim. Biophys. Acta 1426, 297–307. Golgi. Biochim. Biophys. Acta 1426, 309–322.